Comparison of an Extracellular V. Total DNA Extraction Approach for Environmental DNA- Based Monitoring of Sediment Biota

Marine and Freshwater Research © CSIRO 2021 https://doi.org/10.1071/MF20269_AC

Supplementary material

Comparison of an extracellular v. total DNA extraction approach for environmental DNA- based monitoring of sediment biota

Johan PansuA,B,C,D, Michelle B. ChapmanC, Grant C. HoseC and Anthony A. CharitonC

AStation Biologique de Roscoff, UMR 7144, CNRS, Sorbonne Université, Place Georges Teissier, F-29680 Roscoff, France.

BCSIRO Ocean & Atmosphere, New Illawarra Road, Lucas Heights, NSW 2234, Australia.

CDepartment of Biological Sciences, Macquarie University, Balaclava Road, Macquarie Park, NSW 2109, Australia.

DCorresponding author. Present address: Institut des Sciences de l’Évolution de Montpellier (ISEM), CNRS, IPHE, IRD, Université de Montpellier, Place Eugène Bataillon, F-34000 Montpellier, France. Email: [email protected]

Page 1 of 13

Table S1. Statistics regarding metabarcoding data processing

Bacteria (16S) Eukaryote (18S) Metazoan (COI)

Mean number Mean number Mean number Number Number Number Number Number Number of reads per of reads per of reads per of reads of OTUs of reads of OTUs of reads of OTUs sample sample sample

After 3366075 - - 7813456 - - 3051786 - - demultiplexing

Post-GHAP 2620834 27836 36913 +/- 1524 6582313 12663 88950 +/- 1794 2611810 12818 40182 +/- 2201 pipeline

Post-Filtering (before removing 2555517 27093 35993 +/- 1499 6443202 10806 87070 +/- 1753 1674650 5367 26166 +/- 1506 rare OTUs)

Final dataset 1708175 1362 24059 +/- 1026 5672529 998 76655 +/- 1533 1585648 1414 25575 +/- 1358

Table S2. Pairwise comparisons in pre-PCR (polymerase chain reaction) concentrations among the different protocol variants Z-values from Dunn’s multiple-comparison tests are reported here. Significance: *, P < 0.05; **, P < 0.01; ***, P < 0.001. Over the entire dataset (column ‘All’), only pairwise comparisons involving the totDNA extraction protocol with 1 g of sediment are significant

Page 2 of 13 Table S3. Pairwise comparisons in operational taxonomic unit (OTU) richness among the different protocol variants for each primer pair Z-value from the Dunn’s multiple-comparison tests are reported here. None of the Z values was significant (P > 0.05), although marked differences were observed for eukaryotes

Table S4. Results of perMANOVA analyses conducted per sediment type to test for the effect of the extraction type (extDNA and totDNA, regardless of the initial mass of sediment) on Bray–Curtis distances between samples

Page 3 of 13 Table S5. Mean (± s.e.m.) relative read abundance (RRA) of phyla obtained from each protocol variant and averaged fold change (AFC) ratio between totDNA and extDNA extractions Only pairs of protocols with similar amount of starting material were compared using AFC ratios. For each clade, AFC values reflects the positive bias (i.e. enrichment) by either of the extraction method (extDNA or totDNA). Significant differences between the two approaches were assessed using Wilcoxon test. Significance: *, P < 0.05; **, P < 0.01; ***, P < 0.001. Only the 10 most abundant phyla for each marker are represented. Significant differences were observed for eukaryotes and metazoans, but not bacteria

Bacteria (16S)

extDNA 1 g vs totDNA 1 g

Mean (± SEM) RRA Mean (± SEM) RRA Enriching Phylum AFC Significance extDNA 1 g totDNA 1 g method Acidobacteria 6.16% ± 1.25 8.02% ± 0.87 1.30 totDNA 1 g Actinobacteria 2.52% ± 0.72 2.58% ± 0.30 1.02 totDNA 1 g Bacteroidetes 8.88% ± 1.42 10.04% ± 1.89 1.13 totDNA 1 g Chloroflexi 2.72% ± 0.52 3.51% ± 0.82 1.29 totDNA 1 g Euryarchaeota 1.16% ± 0.46 1.09% ± 0.39 1.07 extDNA 1 g Firmicutes 0.47% ± 0.06 0.65% ± 0.08 1.39 totDNA 1 g Planctomycetes 1.09% ± 0.39 0.93% ± 0.32 1.18 extDNA 1 g Proteobacteria 63.52% ± 1.83 59.51% ± 1.44 1.07 extDNA 1 g Thaumarchaeota 2.73% ± 0.93 3.14% ± 1.20 1.15 totDNA 1 g Verrucomicrobia 2.74% ± 0.58 2.48% ± 0.53 1.10 extDNA 1 g

extDNA 10 g vs totDNA 10 g

Mean (± SEM) RRA Mean (± SEM) RRA Enriching Phylum AFC Significance extDNA 10 g totDNA 10 g method Acidobacteria 5.93% ± 1.08 7.58% ± 1.21 1.28 totDNA 10 g Actinobacteria 2.55% ± 0.63 2.25% ± 0.36 1.13 extDNA 10 g Bacteroidetes 8.70% ± 1.35 11.57% ± 1.91 1.33 totDNA 10 g Chloroflexi 2.54% ± 0.46 2.99% ± 0.47 1.18 totDNA 10 g Euryarchaeota 1.32% ± 0.45 0.95% ± 0.35 1.39 extDNA 10 g Firmicutes 0.49% ± 0.07 0.77% ± 0.13 1.56 totDNA 10 g Planctomycetes 1.18% ± 0.39 0.66% ± 0.22 1.80 extDNA 10 g Proteobacteria 62.95% ± 1.60 61.74% ± 1.05 1.02 extDNA 10 g Thaumarchaeota 3.52% ± 1.15 1.77% ± 0.78 1.98 extDNA 10 g Verrucomicrobia 2.27% ± 0.39 2.12% ± 0.36 1.07 extDNA 10 g

extDNA 200 g

Mean (± SEM) RRA Phylum extDNA 200 g Acidobacteria 6.31% ± 1.22 Actinobacteria 1.02% ± 0.28 Bacteroidetes 8.80% ± 1.38 Chloroflexi 1.16% ± 0.21 Euryarchaeota 1.56% ± 0.58 Firmicutes 0.52% ± 0.05 Planctomycetes 1.02% ± 0.33 Proteobacteria 64.95% ± 1.93 Thaumarchaeota 3.35% ± 0.98 Verrucomicrobia 2.32% ± 0.36

Page 4 of 13 Table S5. (Cont.) Eukaryote (18S)

extDNA 1 g vs totDNA 1 g

Mean (± SEM) RRA Mean (± SEM) RRA Enriching Phylum AFC Significance extDNA 1 g totDNA 1 g method Annelida 16.07% ± 4.84 6.78% ± 2.15 2.37 extDNA 1 g Arthropoda 10.93% ± 3.92 21.47% ± 4.38 1.96 totDNA 1 g Bacillariophyta 10.56% ± 2.21 5.64% ± 1.10 1.87 extDNA 1 g Chlorophyta 0.77% ± 0.11 7.94% ± 2.86 10.25 totDNA 1 g Ciliophora 4.70% ± 1.00 2.70% ± 0.35 1.74 extDNA 1 g Gastrotricha 5.77% ± 1.16 4.30% ± 2.45 1.34 extDNA 1 g ** Myzozoa 5.15% ± 1.80 4.81% ± 1.06 1.07 extDNA 1 g Nematoda 3.62% ± 0.97 8.20% ± 1.19 2.27 totDNA 1 g ** Platyhelminthes 21.72% ± 5.20 15.02% ± 3.58 1.45 extDNA 1 g Xenacoelomorpha 4.58% ± 1.68 6.38% ± 2.99 1.40 totDNA 1 g

extDNA 10 g vs totDNA 10 g

Mean (± SEM) RRA Mean (± SEM) RRA Enriching Phylum AFC Significance extDNA 10 g totDNA 10 g method Annelida 15.62% ± 4.22 12.28% ± 4.48 1.27 extDNA 10 g Arthropoda 11.64% ± 3.30 17.19% ± 3.40 1.48 totDNA 10 g Bacillariophyta 9.31% ± 1.82 4.82% ± 0.95 1.93 extDNA 10 g Chlorophyta 0.79% ± 0.17 5.81% ± 2.14 7.31 totDNA 10 g Ciliophora 4.96% ± 1.29 2.93% ± 0.56 1.69 extDNA 10 g Gastrotricha 6.32% ± 1.45 4.76% ± 1.88 1.33 extDNA 10 g Myzozoa 4.90% ± 1.39 4.18% ± 1.07 1.17 extDNA 10 g Nematoda 1.78% ± 0.37 5.56% ± 0.71 3.12 totDNA 10 g *** Platyhelminthes 23.57% ± 5.65 17.75% ± 4.31 1.33 extDNA 10 g Xenacoelomorpha 4.84% ± 1.45 7.98% ± 3.01 1.65 totDNA 10 g

extDNA 200 g

Mean (± SEM) RRA Phylum extDNA 200 g Annelida 18.20% ± 5.74 Arthropoda 9.30% ± 2.92 Bacillariophyta 11.68% ± 2.65 Chlorophyta 0.74% ± 0.15 Ciliophora 4.73% ± 1.27 Gastrotricha 5.43% ± 1.40 Myzozoa 4.69% ± 1.15 Nematoda 0.69% ± 0.11 Platyhelminthes 23.80% ± 6.04 Xenacoelomorpha 5.47% ± 1.68

Page 5 of 13 Table S5. (Cont.)

Metazoan (COI)

extDNA 1 g vs totDNA 1 g

Mean (± SEM) RRA Mean (± SEM) RRA Enriching Phylum AFC Significance extDNA 1 g totDNA 1 g method Annelida 35.21% ± 7.31 22.62% ± 5.25 1.56 extDNA 1 g Arthropoda 15.34% ± 5.73 26.64% ± 3.25 1.74 totDNA 1 g * Cnidaria 2.21% ± 1.03 2.81% ± 0.89 1.27 totDNA 1 g Gastrotricha 0.39% ± 0.07 1.65% ± 0.56 4.24 totDNA 1 g Mollusca 31.41% ± 10.50 22.41% ± 6.63 1.40 extDNA 1 g Nematoda 0.25% ± 0.13 3.16% ± 1.44 12.73 totDNA 1 g *** Platyhelminthes 0.87% ± 0.27 0.93% ± 0.22 1.06 totDNA 1 g Porifera 0.64% ± 0.14 0.76% ± 0.19 1.18 totDNA 1 g Rotifera 0.10% ± 0.06 1.21% ± 0.50 11.64 totDNA 1 g Xenacoelomorpha 3.00% ± 1.1 1.83% ± 0.66 1.64 extDNA 1 g

extDNA 10 g vs totDNA 10 g

Mean (± SEM) RRA Mean (± SEM) RRA Enriching Phylum AFC Significance extDNA 10 g totDNA 10 g method Annelida 45.39% ± 7.70 29.44% ± 6.79 1.54 extDNA 10 g Arthropoda 11.50% ± 3.37 22.41% ± 2.84 1.95 totDNA 10 g * Cnidaria 2.29% ± 0.69 1.49% ± 0.41 1.53 extDNA 10 g Gastrotricha 0.57% ± 0.16 1.69% ± 0.63 2.98 totDNA 10 g Mollusca 23.86% ± 9.60 28.48% ± 8.90 1.19 totDNA 10 g Nematoda 0.28% ± 0.12 0.76% ± 0.25 2.75 totDNA 10 g * Platyhelminthes 0.70% ± 0.26 0.79% ± 0.19 1.13 totDNA 10 g Porifera 0.72% ± 0.21 0.54% ± 0.21 1.32 extDNA 10 g Rotifera 0.14% ± 0.07 0.69% ± 0.28 5.00 totDNA 10 g Xenacoelomorpha 2.77% ± 1.23 1.90% ± 0.66 1.46 extDNA 10 g

extDNA 200 g

Mean (± SEM) RRA Phylum extDNA 200 g Annelida 41.25% ± 8.56 Arthropoda 9.44% ± 3.12 Cnidaria 2.29% ± 0.91 Gastrotricha 0.44% ± 0.16 Mollusca 31.90% ± 10.01 Nematoda 0.16% ± 0.10 Platyhelminthes 0.93% ± 0.33 Porifera 0.73% ± 0.21 Rotifera 0.09% ± 0.06 Xenacoelomorpha 3.63% ± 1.32

Page 6 of 13 Aquatic system

Five samples

Five sub-samples totDNA totDNA extDNA extDNA extDNA (one per extraction 1 g 10 g 1 g 10 g 200 g protocol)

Three PCR amplifications 18S COI 16S eukaryote metazoa bacteria

Fig. S1. Study design. Three aquatic systems (pond, estuarine and marine) were studied; five samples (~500 g) were collected in each (sample n = 15). From each sample, five subsamples of different volume were used for assessing the efficiency of two DNA extraction approaches: 1 g, 10 g and 200 g for the extracellular DNA method, 1 g and 10 g for the total DNA extraction method (extraction n = 75). Each DNA extract was then amplified with three different primer pairs to characterise bacterial, eukaryote and metazoan communities (PCR n = 225).

Page 7 of 13 25 extDNA 1 g totDNA 10 g PPondond sediment A. 16S primers (Bacteria) extDNA 200 g

) totDNA 1 g EstuarEstuarineine sediment -1 extDNA 10 g MarMarineine sediment µL 20

5 Post-PCR DNA concentration (ng 0

) B. 18S primers (Eukaryote) -1 µL 20

5 Post-PCR DNA concentration (ng 0

25 C. COI primers (Metazoan) ) -1 µL 20

5 Post-PCR DNA concentration (ng 0 extDNA 1 g extDNA totDNA 1 g 10 g extDNA totDNA 10 g 200 g extDNA 1 g extDNA totDNA 1 g 10 g extDNA totDNA 10 g extDNA 200 g 1 g extDNA totDNA 1 g 10 g extDNA totDNA 10 g 200 g extDNA

Fig. S2. Mean (± s.e.m.) post-PCR (polymerase chain reaction ) concentrations per extraction protocol variant for each sediment type. Each row corresponds to one of the primer pairs used in this study. (a) 16S bacteria, (b) 18S eukaryote, (c) COI metazoan.

Page 8 of 13 Fig. S3. Operational taxonomic unit (OTU) accumulation curve per sediment type (pond, estuarine or marine) according to the extraction protocol variant for bacteria (top panel), eukaryotes (central panel) and metazoans (bottom panel). Colours correspond to the extraction protocol variant. Estuary samples (plain line); marine samples (dashed line); pond samples (dotted line). Coleman method for species accumulation curve was used to estimate the expected richness. Errors bars correspond to the standard deviation from 100 permutations.

Page 9 of 13

Bacteria (16S) Eukaryote (18S) Metazoa (COI)

. 7

F = 1.03, df = 4, P = 0.42 F = 0.29, df = 4, P = 0.88 . F = 3.41, df = 4, P < 0.04

0 0

6 1

F = 3.87, df = 4, P < 0.02 0

. F = 2.06, df = 4, P = 0.12 F = 3.18, df = 4, P < 0.04

y .

e 0

2 5

1 2

. .

0 0

D 0

. 0

extDNA extDNA extDNA totDNA totDNA

F = 3.69, df = 4, P < 0.03 0 F = 1.52, df = 4, P = 0.24

6 .

0 1 g 10 g 200 g 1 g 10 g

o 0 r

5 Extraction protocol

. 0 extDNA extDNA extDNA totDNA totDNA extDNA extDNA extDNA totDNA totDNA 1 g 10 g 200 g 1 g 10 g 1 g 10 g 200 g 1 g 10 g Extraction protocol Extraction protocol

Fig. S4. Homogeneity of samples according to the extraction protocol variant. For each sediment type and taxonomic group, samples are grouped per extraction protocol variant, and Bray–Curtis distances of each sample to the centroid of its group are reported (‘betadisper’ analysis; J. Oksanen, F. G. Blanchet, M. Friendly, R. Kindt, P. Legendre, D. McGlinn, P. R. Minchin, R. B. O’Hara, G. L. Simpson, P. Solymos, M. H. H. Stevens, E. Szoecs, and H. Wagner, https://CRAN.R- project.org/package=vegan). Low distances to the centroid indicate high homogeneity among sample replicates. ANOVA was performed for each taxon in each sediment type to test for significant differences between protocols.

Page 10 of 13

Fig. S5. Comparison of phylum average relative abundances between the totDNA and extDNA extraction approaches for bacteria, eukaryotes and metazoans. The solid line indicates a 1 : 1 relationship. Data presented here are log-transformed, one read have been added to each operational taxonomic unit (OTU) count prior analysis to avoid null values.

Page 11 of 13

Code for the inertia decomposition analysis The inertia of a group of points represents their dispersion and corresponds to the sum of the squared distances from each point to centroid of the group (i.e. cloud of points). The additive property of the inertia makes that the inertia of a group of points is equal to the sum of all the inertia of subgroups (Prud’Homme 2012). Here, for each sedimentary habitat, five samples were collected, and each of them was extracted using five different protocols. The variation between DNA extracts from the same habitat occurs at two different levels, namely, variation owing to sampling and variation owing to extraction. We used the additive property of the inertia to decompose the total inertia (Itot) between the portion owing to the variation between extractions from a sample (Iext) and the portion owing to the biological variation between samples (Isamp): Itot = Isamp + Iext

For this analysis, we first conducted correspondence analyses for each sediment type, by retaining the highest possible number of dimensions D. We then calculated the total inertia (i.e. the dispersion of the entire cloud of points) by summing the squared distances of each point to the centroid of these points in the D dimensions. Points were then grouped according to their sample of origin, and the centroid of each of these subgroups was calculated. The inertia owing to the variation between extractions corresponds then to the sum of the squared distances of points to their respective subgroups. The fraction of the inertia owing to the variation between extractions from a same sample FIext (i.e. ‘extraction effect’) was then defined as FIext = Iext / Itot. Finally, because we have only two levels of variation here, the fraction of the inertia owing to the variation between samples is FIsamp is: FIsamp = 1 – FIext.

We thank Eric Coissac and Sophie Prud’Homme for sharing their code, which was originally developed as part of Sophie Prud’Homme’s Master of Research thesis (Prud’Homme 2012).

###Function ‘barycentre’ for calculating the centroid of a group of points barycentre = function(line,grp) {

aggmean = function(x) aggregate(x,grp,mean)[,2] grp=as.character(grp) weight = table(grp) grp = list(class=grp) centre = rbind(apply(line,2,aggmean)) names=aggregate(as.vector(line[,1]),grp,mean)[,1] if (length(names)==1)

rownames(centre)[1]=names else

rownames(centre)=names attr(centre,"weight")=weight centre

Page 12 of 13

}

### Function ‘inertia’ to calculate the inertia of a group of points inertia = function(line,grp) {

centre = barycentre(line,grp)

(line - centre[grp,])**2

}

library(ade4)

###Correspondance analysis

#Tab: mOTUs x sample matrix coa = dudi.coa(Tab, scan = FALSE, nf = nrow(Tab)-1)

###Total Inertia

#nbSamp: total number of points included in the CA

Itot = sum(inertia(coa$li, rep(1, nbSamp)))

###Inertia extraction

#SubG: vector describing the sub-group (i.e. the original sample) of each point (e.g. S1, S1, S2, S2)

Iext: sum(inertie(pca1$li, SubG))

###Fraction of the inertia due to the variation between extractions

FIext = Iext / Itot

###Fraction of the inertia due to the variation between sample

FIsamp = 1- FIext

Reference Prud’Homme, S. (2012) Environmental DNA metabarcoding for plant biodiversity assessment. M.Res. Thesis, Ecole Normale Supérieure Lyon & University of Lyon, France.

Page 13 of 13